American economic historians generally believe US modern economic growth began before 1840, was broadly based (not driven by a single activity such as cotton cultivation or the railroad), and the acceleration was gradual (rather than a sudden takeoff). These claims are treated as conjectures rather than solid conclusions because the census did not publish good economic data before 1840. But the earlier period was not a "statistical dark age," as sometimes asserted. From 1817 on, systematic statistics on postal activity are available at the local level at high frequencies. This paper assembles these data and asks what we can learn.